Design and Analysis of LoS MIMO Systems with Uniform Circular Arrays

TL;DR

This paper develops a comprehensive model and optimal design strategies for LoS MIMO systems with uniform circular arrays, accounting for misalignments, and demonstrates near-orthogonal channel matrices enabling simple decoding.

Contribution

It introduces a generic channel model for UCA-based LoS MIMO systems considering misalignments and proposes an optimal design method based on RPDR to maximize capacity.

Findings

Singular values are independent of tilting and center-shift.

RPDR is a key parameter for system design.

Optimal UCA design yields near-orthogonal channels.

Abstract

We consider the design of a uniform circular array (UCA) based multiple-input multiple-output (MIMO) system over line-of-sight (LoS) environments in which array misalignment exists. In particular, optimal antenna placement in UCAs and transceiver architectures to achieve the maximum channel capacity without the knowledge of misalignment components are presented. To this end, we first derive a generic channel model of UCA-based LoS MIMO systems in which three misalignment factors including relative array rotation, tilting and center-shift are reflected concurrently. By factorizing the channel matrix into the singular value decomposition (SVD) form, we demonstrate that the singular values of UCA-based LoS MIMO systems are \textit{independent} of tilting and center-shift. Rather, they can be expressed as a function of the \textit{radii product-to-distance ratio} (RPDR) and the angle of relative array rotation. Numerical analyses of singular values show that the RPDR is a key design parameter of UCA systems. Based on this result, we propose an optimal design method for UCA systems which performs a one-dimensional search of RPDR to maximize channel capacity. It is observed that the channel matrix of the optimally designed UCA system is close to an orthogonal matrix; this fact allows channel capacity to be achieved by a simple zero-forcing (ZF) receiver. Additionally, we propose a low-complexity precoding scheme for UCA systems in which the optimal design criteria cannot be fulfilled because of limits on array size. The simulation results demonstrate the validity of the proposed design method and transceiver architectures.